Skip to content

Improve KafkaReadStream control #75

Open
@vietj

Description

@vietj

Currently the KafkaReadStream delivers item by batch of 10 (this should be configurable btw). When doing seek operations, the stream needs to be paused otherwise the handler will likely receive unwanted records that have been fetch and are delivered by the batch. This seems due to the fact that the current iterator is accessed from the worker thread only, unlike the paused flag.

One way to achieve it might be to pause() the consumer in seek operations. Alternatively the current iterator could be modified outside of the worker thread and checked when delivering items. Setting it to null would mean that a seek operation was required.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions